Difference between revisions of "ToolMatch"

From Earth Science Information Partners (ESIP)
 
(6 intermediate revisions by one other user not shown)
Line 2: Line 2:
 
* [https://docs.google.com/spreadsheets/d/1XiCun7yIaGj3v5eXzG8yXAjAha7tAcZZT4tTnpguPEU/edit#gid=403447662 Glossary of Terms] (gdoc spreadsheet)
 
* [https://docs.google.com/spreadsheets/d/1XiCun7yIaGj3v5eXzG8yXAjAha7tAcZZT4tTnpguPEU/edit#gid=403447662 Glossary of Terms] (gdoc spreadsheet)
 
* [https://docs.google.com/spreadsheets/d/1tHZPZ4xtkpzEWR_l5T1t7muOqmtLJ4EBZnID3lPKtIA/edit#gid=72976985 Data Collections and Tools Characteristics] (gdoc spreadsheet)
 
* [https://docs.google.com/spreadsheets/d/1tHZPZ4xtkpzEWR_l5T1t7muOqmtLJ4EBZnID3lPKtIA/edit#gid=72976985 Data Collections and Tools Characteristics] (gdoc spreadsheet)
* [https://toolmatch.hackpad.com/Glossary-Ontology-Questions-TNQPPvUmcWg Glossary/Ontology questions]
+
* [https://toolmatch.hackpad.com/Glossary-Ontology-Questions-TNQPPvUmcWg Glossary/Ontology questions] (hackpad)
 +
* [https://toolmatch.hackpad.com/Issues-with-Data-Collection-Tool-Identification-oiELoplmv2s Issues with data collection and tool identification] (discussions pieced together on hackpad)
 
* [[ToolMatch To Do]]
 
* [[ToolMatch To Do]]
* [[ToolMatch To Do]]<br />
+
* Working meeting recordings are hosted on the [https://github.com/ESIPFed/toolmatch-ontology ToolMatch Ontology Github repository].
 +
:: To download, go to the link and click the '''Raw''' button.
 +
:: The [https://github.com/ESIPFed/toolmatch-ontology/blob/master/meetings/Toolmatch_20160323.mp4 March 23, 2016] recording includes the most recent discussion (primarily) focused on term definitions.
 +
:: [https://github.com/ESIPFed/toolmatch-ontology/blob/master/meetings/Toolmatch_20160328.mp4 March 28, 2016] recording, term definitions and possible ontology alignment.
 
<br />
 
<br />
 
'''previous'''
 
'''previous'''
Line 59: Line 63:
 
* TBD<br />
 
* TBD<br />
 
<br />
 
<br />
'''2013'''
+
'''2013'''<br />
 
Per ESIP Winter 2013 Meeting:
 
Per ESIP Winter 2013 Meeting:
 
# Continue identifying tools that could be included and describing pertinent characteristics
 
# Continue identifying tools that could be included and describing pertinent characteristics
Line 128: Line 132:
 
* [[User:Nhoebelheinrich|Nancy Hoebelheinrich]] 11:51, 26 July 2013 (PST)
 
* [[User:Nhoebelheinrich|Nancy Hoebelheinrich]] 11:51, 26 July 2013 (PST)
 
* [[User:Brandonnodnarb|Brandon Whitehead]] 16:27, 6 June 2015 (UTC)
 
* [[User:Brandonnodnarb|Brandon Whitehead]] 16:27, 6 June 2015 (UTC)
 
= Working Documents 2016 =
 
 
== Meeting Recordings ==
 
 
These are hosted in the Toolmatch Ontology Github [https://github.com/ESIPFed/toolmatch-ontology repository]. To download, go to the link and click the '''Raw''' button.
 
 
[https://github.com/ESIPFed/toolmatch-ontology/blob/master/meetings/Toolmatch_20160323.mp4 March 23, 2016]
 

Latest revision as of 15:49, March 28, 2016

What's New?

To download, go to the link and click the Raw button.
The March 23, 2016 recording includes the most recent discussion (primarily) focused on term definitions.
March 28, 2016 recording, term definitions and possible ontology alignment.


previous

Problem Statement and Use Case

For a given dataset, it is difficult to find the tools that can be used to work with the dataset. In many cases, the information that Tool A works with Dataset B is somewhere on the Web, but not in a readily identifiable or discoverable form. In other cases, particularly more generalized tools, the information does not exist at all, until somebody tries to use the tool on a given dataset.

Thus, the simplest, most prevalent use case is for a user to search for the tools that can be used with a given dataset. A further refinement would be to specify what the tool can do with the dataset, e.g., read, visualize, map, analyze, reformat.

  • Note that the Energy Cluster is actually looking for this kind of tool. (See Rahul.)

Proposed Solution

Often, whether a tool is likely to work with a dataset can be inferred through simple rules. For example, knowing that a data is available in netcdf/CF1 and is on a lat/long grid is typically sufficient to infer the data can be viewed through Panoply. Secondly, the problem lends itself to crowdsourcing: once one user has found a given tool to be usable with a given dataset, this holds true for all users, and so the information should be promulgated.

We propose the construction of RDF triples that record the fact that a tool works with a particular dataset. It would be based on a simple ontology, with minimal information about the dataset (enough to uniquely identify it and present it as an option in a user interface). There would be slightly more information captured for the tool. A simple user interface would allow a user to select a dataset or paste in a unique dataset identifier.

Requirements

  • Tools can be either downloadable tools or online services
  • Datasets should be identifiable either through GCMD DIF ID or DOI.
  • Reformatted data and reformatting services (WCS, OPeNDAP) should be considered in compatibility.
  • A simple User interface should provide the ability to search for tools compatible with a certain dataset
  • Users should be able to see a brief description of the tool.
  • Users should be presented with a website for the tool in the search results.
  • A simple mechanism should be provided for authoring the RDF links. (This does not mean they will author RDF directly in, say, TTL or XML. They might do so through simple hashtags).
  • The "system" should be able to do some simple inferencing.

Artifacts

Random Notes

Relationship to Servicecasting

This could be an underlying infrastructure for generating service casts...

Inferencing

There are three kinds of inferences that can greatly cut down the cost of authoring the RDF.

  1. For tools based on netcdf-java, the likelihood of usability can often be inferred from availability in netCDF or OPeNDAP and presence of CF-1 coordinates. Note that this is not an ironclad guarantee however.
  2. Many data collections have "homeomorphic" siblings: datasets with the same or similar variables and data format. For example, AIRH2RET, AIRS2RET and AIRX2RET are AIRS Level 2 datasets that all have the same variables (roughly) and data structure and format. Usability for one in a given tool strongly implies usability for a sibling dataset.
  3. Some tools by their nature are NOT suited to an entire class of datasets, such as GrADS for mapping Level 2 data.

Tool Scripts

  1. Some tools are actually scripts written to work within other tools. An example of this is the opengrads cookbook, which contains scripts for reading and working with certain data collections in the GrADS tool.

Implementation Plan

2016

  • TBD


2013
Per ESIP Winter 2013 Meeting:

  1. Continue identifying tools that could be included and describing pertinent characteristics
  2. Identify collections that could be used as test cases
  3. Adjust SADL and CMAP representations of ToolMatch models as they evolve
  4. Coordinate with ESIP Energy Cluster as they are working on similar efforts
  5. Identify next steps in joint telecons with the Energy Cluster & other interested parties
  6. Create a plan for realistic implementation to be demonstrated at ESIP Summer mtg 2014

Presentations

Meetings

Meetings from June 2015 to present are documented on the ToolMatch Hackpad
Call in information
We are now using GoToMeeting

  • You may also dial in using your phone: +1 (872) 240-3212
  • if prompted, the access code is: 156-154-397


2015

2014

2013

SIGNUP SHEET

Make a spot for you to leave your "~~~~" if you want to contribute to this collaboration.: